skip to content
Site header image reelikklemind

🎙️ The Knowledge Project: Benedict Evans

The Patterns That Everyone Else Misses

🌲 Evergreen 🎧 Podcast 👔 Business 📯 History

🎙️ The Knowledge Project: Benedict Evans, The Patterns That Everyone Else Misses

PODCAST INFORMATION

The Knowledge Project
Benedict Evans: The Patterns That Everyone Else Misses
Shane Parrish (Host)
Benedict Evans (Guest) - Technology analyst known for his insightful takes on platform shifts in the tech industry)
Episode Duration: Approximately 1 hour and 11 minutes

🎧 Listen here.


HOOK

Benedict Evans reveals how AI represents not an unprecedented revolution, but rather another platform shift that follows historical patterns, challenging both the hype and fear surrounding today's technological transformation.


ONE-SENTENCE TAKEAWAY

AI represents the most significant platform shift since the iPhone, but like all technological revolutions, its true impact will unfold in ways we cannot fully predict, with incumbents facing disruption while new opportunities emerge for those who understand the patterns of technological change.


SUMMARY

In this episode of The Knowledge Project, host Shane Parrish engages in a wide-ranging conversation with technology analyst Benedict Evans about AI, platform shifts, and the patterns that define technological revolutions. Evans, known for his ability to spot patterns others miss, offers a nuanced perspective on AI that positions it as "the biggest thing since the iPhone" but cautions against viewing it as something more transformative like the industrial revolution or electricity.

The conversation begins with Evans establishing his "centrist" view on AI, which he considers controversial in a field dominated by extreme positions. He argues that while AI will certainly transform industries, employment, and productivity over the next 10-15 years, it will eventually become "just software" like previous technological innovations.

Evans provides historical context by examining previous platform shifts, including the internet revolution of the 1990s and the mobile internet transformation. He shares fascinating examples of how uncertain these shifts were at their inception, noting that in 1995, analysts were not sure if the internet would be decentralized or controlled by media companies, or whether email users would outnumber web users. Similarly, he recounts how the mobile internet's evolution was unpredictable, with few anticipating that smartphones would essentially become "small Macs" that would eventually replace PCs as the center of the tech industry.

A significant portion of the discussion explores how incumbents fare during platform shifts. Evans challenges the notion that having data necessarily provides an advantage to established companies, pointing out that people thought Microsoft would dominate the internet, IBM would win in PCs, and Meta would lead in mobile—all predictions that proved incorrect. He uses the Kodak example to illustrate how even when incumbents recognize technological shifts and attempt to adapt, they may still fail due to fundamental changes in business models and market dynamics.


Regarding AI specifically, Evans makes several counterintuitive arguments. He suggests that contrary to popular belief, data may not provide a significant advantage to companies like Google or Meta because large language models require such enormous amounts of generalized text that everyone essentially has access to the same data. He also questions whether current AI models can truly differentiate themselves from one another, suggesting that in a double-blind test, most people would not be able to distinguish between outputs from Grock, Claude, Gemini, Mistral, or DeepSeek.

The conversation touches on numerous other topics, including the challenges of regulating AI at the right level of abstraction, the economic trade-offs inherent in regulation, how people use and perceive AI tools (with survey data showing only about 10% of people use AI tools daily), the value of original thinking in an age of AI-generated content, advice for students and professionals navigating technological change, and analysis of how major tech companies (Apple, Google, Microsoft, Meta, Amazon) are positioned for the AI era.

Throughout the episode, Evans demonstrates his ability to identify patterns across technological revolutions while acknowledging the uncertainties inherent in predicting how new technologies will evolve. His perspective offers a valuable counterbalance to both the hype and fear surrounding AI, emphasizing the importance of understanding historical context when evaluating technological change.


INSIGHTS

  1. AI represents another platform shift rather than an unprecedented revolution, following patterns established by previous technological transformations like the internet and mobile computing.
  2. Despite the hype, AI's impact on employment and the economy will likely resemble previous platform shifts rather than representing something fundamentally different.
  3. Incumbents rarely maintain their advantage during platform shifts, even when they recognize the change and attempt to adapt, as seen with Microsoft in internet, IBM in PCs, and Kodak in digital photography.
  4. Data may not provide the competitive advantage in AI that many assume, as large language models require such enormous amounts of generalized text that everyone essentially has access to the same data.
  5. Current AI models may be more similar than different, with branding and distribution potentially mattering more than underlying technical differences.
  6. Regulation of AI requires careful consideration of trade-offs, as overly restrictive approaches could stifle innovation while failing to address actual harms.
  7. AI adoption patterns show that only about 10% of people use AI tools daily, suggesting we are still in early stages of finding compelling use cases for the general population.
  8. Original thinking and asking the right questions will become increasingly valuable as AI becomes better at generating conventional content and analysis.
  9. The true impact of technological shifts often takes years to materialize and rarely unfolds in the ways experts initially predict.
  10. Major tech companies face different challenges in the AI era, with Apple potentially being "Microsofted" as smartphones become commoditized devices for accessing AI services in the cloud.


FRAMEWORKS & MODELS

  1. Platform Shift Analysis Framework
    • Components: Historical context, incumbent behavior, value capture points, adoption patterns, and long-term evolution
    • How it works: Examines how major technological transformations follow similar patterns despite surface differences
    • Application: Evans uses this framework to analyze AI by comparing it to previous shifts like the internet, mobile computing, and digital photography
    • Significance: Helps avoid both overhyping and underestimating new technologies by placing them in historical context
    • Example: Evans compares AI to the internet revolution of the 1990s, noting how uncertain the internet's eventual form was at its inception
  2. Incumbent Response Model
    • Components: Absorption attempt, unbundling of existing businesses, contingent factors, and market structure changes
    • How it works: Explains how established companies typically respond to technological disruptions and why they often fail despite recognizing the change
    • Application: Used to analyze how companies like Google, Apple, and Microsoft might navigate the AI transition
    • Significance: Provides a template for understanding why technological disruptions consistently challenge incumbents
    • Example: Detailed analysis of Kodak's response to digital photography, showing how even successful adaptation to the new technology failed to save the business
  3. AI Value Capture Framework
    • Components: Model differentiation, distribution advantages, network effects potential, and integration with existing products
    • How it works: Analyzes where value might be captured in the AI ecosystem, from foundational models to applications
    • Application: Used to evaluate the competitive positions of companies like OpenAI, Google, Microsoft, and Meta
    • Significance: Helps cut through the hype to identify sustainable business models in the AI space
    • Example: Discussion of why ChatGPT dominates usage despite similar capabilities to other models, emphasizing the importance of branding and distribution


QUOTES

  1. "My sort of base case is to say this is kind of another platform shift and all the new stuff will be built around this for the next 10 or 15 years and then there will be something else and so the impact on employment will be kind of like the impact on employment from the other platform shifts." - Benedict Evans
    • Context: Early in the conversation when establishing his centrist view on AI
    • Significance: Captures Evans' core thesis that AI represents significant change but follows historical patterns rather than being unprecedented
  2. "The very high level threat to Google is that you have this moment of discontinuity in which everybody resets their priority and reconsiders their defaults. And so it is no longer just the default that you go and use Google." - Benedict Evans
    • Context: Discussing how platform shifts challenge incumbents by changing user behavior patterns
    • Significance: Explains why even dominant companies like Google face vulnerability during technological transitions
  3. "It seems to me right now you could do like a double blind test of the same prompt given to Grock, Claude, Gemini, Mistral, Deep Seek. I bet most people would not be able to tell which is which." - Benedict Evans
    • Context: Analyzing the differentiation between AI models
    • Significance: Challenges the notion that technical superiority in AI models necessarily translates to user-perceivable differences or competitive advantage
  4. "If you make it really hard and expensive to build houses, houses will be more expensive. You have made that choice. If you do that, you cannot then complain that houses are more expensive." - Benedict Evans
    • Context: Discussing the trade-offs inherent in regulating AI
    • Significance: Illustrates Evans' economic perspective on regulation, emphasizing that policy choices have consequences that must be accepted
  5. "I write something and I think is that I would what I would always would ask in the past is kind of your point about pattern recognition is I look at something and say am I adding value here am I saying something useful am I saying something different am I asking the key question am I pushing further am I pushing the am I asking the next question rather than just answering the obvious questions now I can just say is this what chat GPT would have said and if the answer is this is what chat GPT would have said then I did not publish it." - Benedict Evans
    • Context: Explaining how he uses AI as a benchmark for original thinking
    • Significance: Demonstrates a practical approach to maintaining value creation in an age of AI-generated content


HABITS

  1. Historical Pattern Recognition
    • Practice: Study previous technological shifts to identify common patterns and outcomes
    • Implementation: Regularly examine how technologies like the internet, mobile computing, and digital photography evolved
    • Application: Apply historical insights to evaluate current technological developments like AI
    • Benefit: Avoids both overhyping and underestimating new technologies by placing them in context
    • Pitfall to avoid: Assuming that history repeats exactly rather than rhyming
  2. Questioning Conventional Wisdom
    • Practice: Challenge widely accepted narratives about technological change
    • Implementation: Develop counterarguments to popular positions on technology's impact
    • Application: Apply critical thinking to claims about AI's revolutionary nature or inevitable effects
    • Benefit: Develops more nuanced understanding of technological change and its implications
    • Pitfall to avoid: Contrarianism for its own sake rather than based on evidence
  3. First-Principles Analysis
    • Practice: Break down complex technological phenomena to fundamental components
    • Implementation: Ask basic questions about how technologies work, who benefits, and what changes
    • Application: Analyze AI by examining data requirements, model capabilities, and value capture points
    • Benefit: Cuts through hype to understand what is truly novel versus what follows established patterns
    • Pitfall to avoid: Oversimplifying complex systems by missing important interactions
  4. Multi-Perspective Evaluation
    • Practice: Consider technological change from multiple viewpoints (economic, social, technical, historical)
    • Implementation: Analyze AI through different lenses like regulation, business models, user behavior, and technical capabilities
    • Application: Develop comprehensive understanding of AI's potential impact across domains
    • Benefit: Avoids one-dimensional analysis that misses important implications
    • Pitfall to avoid: Paralysis by analysis through considering too many perspectives without synthesis
  5. Continuous Curiosity
    • Practice: Maintain ongoing interest in technological developments and their implications
    • Implementation: Regularly explore new technologies, read widely, and engage with diverse viewpoints
    • Application: Stay informed about AI developments while maintaining historical perspective
    • Benefit: Develops the pattern recognition skills necessary to identify meaningful technological shifts
    • Pitfall to avoid: Superficial engagement with many topics rather than deep understanding


REFERENCES

  1. Historical Platform Shifts
    • Internet revolution of the 1990s, including the uncertainty about its eventual form
    • Mobile internet transformation and the evolution from "feature phones" to smartphones
    • Digital photography disruption and Kodak's response
    • Personal computer revolution and IBM's initial dominance followed by loss of control
  2. Research and Data Sources
    • Survey data on AI usage patterns showing approximately 10% daily usage
    • Mary Meeker's 1995 internet report with separate forecasts for web users and email users
    • 1995 research firm diagram of "cyberspace" before the internet's form was clear
    • Google TAC trial evidence showing Google's continued search engine superiority
  3. Economic and Regulatory Frameworks
    • Hayek's insights on pricing as an information system and signal of demand
    • "Red Plenty" book about Soviet central planning in the 1960s-1980s
    • Economic principles of regulation having costs and consequences
    • Trade-off analysis in policy making
  4. Business and Technology Analysis
    • Kodak's annual reports from their digital camera era
    • Analysis of browser innovation over the past 25 years
    • Social media platform evolution and network effects
    • Cloud computing business models and their application to AI
  5. Philosophical and Conceptual References
    • "How to Talk About Books You Have not Read" by Pierre Bayard
    • "Bonnheur des Dames" (The Ladies' Delight) by Émile Zola about department store creation
    • Bismarck's concept of the "great man" who recognizes historical moments
    • Paradox of the monkeys and typewriters (infinite library concept)



Crepi il lupo! 🐺